^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1) // SPDX-License-Identifier: GPL-2.0-only
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3) * linux/mm/mmu_notifier.c
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 4) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 5) * Copyright (C) 2008 Qumranet, Inc.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 6) * Copyright (C) 2008 SGI
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 7) * Christoph Lameter <cl@linux.com>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 8) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 9)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 10) #include <linux/rculist.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 11) #include <linux/mmu_notifier.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 12) #include <linux/export.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 13) #include <linux/mm.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 14) #include <linux/err.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 15) #include <linux/interval_tree.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 16) #include <linux/srcu.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 17) #include <linux/rcupdate.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 18) #include <linux/sched.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 19) #include <linux/sched/mm.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 20) #include <linux/slab.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 21)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 22) /* global SRCU for all MMs */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 23) DEFINE_STATIC_SRCU(srcu);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 24)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 25) #ifdef CONFIG_LOCKDEP
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 26) struct lockdep_map __mmu_notifier_invalidate_range_start_map = {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 27) .name = "mmu_notifier_invalidate_range_start"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 28) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 29) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 30)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 31) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 32) * The mmu_notifier_subscriptions structure is allocated and installed in
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 33) * mm->notifier_subscriptions inside the mm_take_all_locks() protected
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 34) * critical section and it's released only when mm_count reaches zero
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 35) * in mmdrop().
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 36) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 37) struct mmu_notifier_subscriptions {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 38) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 39) * WARNING: hdr should be the first member of this structure
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 40) * so that it can be typecasted into mmu_notifier_subscriptions_hdr.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 41) * This is required to avoid KMI CRC breakage.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 42) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 43) struct mmu_notifier_subscriptions_hdr hdr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 44) /* all mmu notifiers registered in this mm are queued in this list */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 45) struct hlist_head list;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 46) bool has_itree;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 47) /* to serialize the list modifications and hlist_unhashed */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 48) spinlock_t lock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 49) unsigned long invalidate_seq;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 50) unsigned long active_invalidate_ranges;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 51) struct rb_root_cached itree;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 52) wait_queue_head_t wq;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 53) struct hlist_head deferred_list;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 54) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 55)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 56) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 57) * This is a collision-retry read-side/write-side 'lock', a lot like a
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 58) * seqcount, however this allows multiple write-sides to hold it at
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 59) * once. Conceptually the write side is protecting the values of the PTEs in
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 60) * this mm, such that PTES cannot be read into SPTEs (shadow PTEs) while any
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 61) * writer exists.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 62) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 63) * Note that the core mm creates nested invalidate_range_start()/end() regions
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 64) * within the same thread, and runs invalidate_range_start()/end() in parallel
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 65) * on multiple CPUs. This is designed to not reduce concurrency or block
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 66) * progress on the mm side.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 67) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 68) * As a secondary function, holding the full write side also serves to prevent
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 69) * writers for the itree, this is an optimization to avoid extra locking
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 70) * during invalidate_range_start/end notifiers.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 71) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 72) * The write side has two states, fully excluded:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 73) * - mm->active_invalidate_ranges != 0
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 74) * - subscriptions->invalidate_seq & 1 == True (odd)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 75) * - some range on the mm_struct is being invalidated
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 76) * - the itree is not allowed to change
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 77) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 78) * And partially excluded:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 79) * - mm->active_invalidate_ranges != 0
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 80) * - subscriptions->invalidate_seq & 1 == False (even)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 81) * - some range on the mm_struct is being invalidated
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 82) * - the itree is allowed to change
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 83) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 84) * Operations on notifier_subscriptions->invalidate_seq (under spinlock):
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 85) * seq |= 1 # Begin writing
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 86) * seq++ # Release the writing state
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 87) * seq & 1 # True if a writer exists
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 88) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 89) * The later state avoids some expensive work on inv_end in the common case of
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 90) * no mmu_interval_notifier monitoring the VA.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 91) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 92) static bool
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 93) mn_itree_is_invalidating(struct mmu_notifier_subscriptions *subscriptions)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 94) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 95) lockdep_assert_held(&subscriptions->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 96) return subscriptions->invalidate_seq & 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 97) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 98)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 99) static struct mmu_interval_notifier *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 100) mn_itree_inv_start_range(struct mmu_notifier_subscriptions *subscriptions,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 101) const struct mmu_notifier_range *range,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 102) unsigned long *seq)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 103) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 104) struct interval_tree_node *node;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 105) struct mmu_interval_notifier *res = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 106)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 107) spin_lock(&subscriptions->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 108) subscriptions->active_invalidate_ranges++;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 109) node = interval_tree_iter_first(&subscriptions->itree, range->start,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 110) range->end - 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 111) if (node) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 112) subscriptions->invalidate_seq |= 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 113) res = container_of(node, struct mmu_interval_notifier,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 114) interval_tree);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 115) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 116)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 117) *seq = subscriptions->invalidate_seq;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 118) spin_unlock(&subscriptions->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 119) return res;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 120) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 121)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 122) static struct mmu_interval_notifier *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 123) mn_itree_inv_next(struct mmu_interval_notifier *interval_sub,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 124) const struct mmu_notifier_range *range)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 125) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 126) struct interval_tree_node *node;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 127)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 128) node = interval_tree_iter_next(&interval_sub->interval_tree,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 129) range->start, range->end - 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 130) if (!node)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 131) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 132) return container_of(node, struct mmu_interval_notifier, interval_tree);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 133) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 134)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 135) static void mn_itree_inv_end(struct mmu_notifier_subscriptions *subscriptions)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 136) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 137) struct mmu_interval_notifier *interval_sub;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 138) struct hlist_node *next;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 139)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 140) spin_lock(&subscriptions->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 141) if (--subscriptions->active_invalidate_ranges ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 142) !mn_itree_is_invalidating(subscriptions)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 143) spin_unlock(&subscriptions->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 144) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 145) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 146)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 147) /* Make invalidate_seq even */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 148) subscriptions->invalidate_seq++;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 149)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 150) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 151) * The inv_end incorporates a deferred mechanism like rtnl_unlock().
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 152) * Adds and removes are queued until the final inv_end happens then
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 153) * they are progressed. This arrangement for tree updates is used to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 154) * avoid using a blocking lock during invalidate_range_start.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 155) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 156) hlist_for_each_entry_safe(interval_sub, next,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 157) &subscriptions->deferred_list,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 158) deferred_item) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 159) if (RB_EMPTY_NODE(&interval_sub->interval_tree.rb))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 160) interval_tree_insert(&interval_sub->interval_tree,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 161) &subscriptions->itree);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 162) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 163) interval_tree_remove(&interval_sub->interval_tree,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 164) &subscriptions->itree);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 165) hlist_del(&interval_sub->deferred_item);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 166) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 167) spin_unlock(&subscriptions->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 168)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 169) wake_up_all(&subscriptions->wq);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 170) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 171)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 172) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 173) * mmu_interval_read_begin - Begin a read side critical section against a VA
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 174) * range
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 175) * @interval_sub: The interval subscription
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 176) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 177) * mmu_iterval_read_begin()/mmu_iterval_read_retry() implement a
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 178) * collision-retry scheme similar to seqcount for the VA range under
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 179) * subscription. If the mm invokes invalidation during the critical section
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 180) * then mmu_interval_read_retry() will return true.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 181) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 182) * This is useful to obtain shadow PTEs where teardown or setup of the SPTEs
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 183) * require a blocking context. The critical region formed by this can sleep,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 184) * and the required 'user_lock' can also be a sleeping lock.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 185) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 186) * The caller is required to provide a 'user_lock' to serialize both teardown
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 187) * and setup.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 188) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 189) * The return value should be passed to mmu_interval_read_retry().
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 190) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 191) unsigned long
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 192) mmu_interval_read_begin(struct mmu_interval_notifier *interval_sub)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 193) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 194) struct mmu_notifier_subscriptions *subscriptions =
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 195) interval_sub->mm->notifier_subscriptions;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 196) unsigned long seq;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 197) bool is_invalidating;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 198)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 199) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 200) * If the subscription has a different seq value under the user_lock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 201) * than we started with then it has collided.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 202) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 203) * If the subscription currently has the same seq value as the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 204) * subscriptions seq, then it is currently between
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 205) * invalidate_start/end and is colliding.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 206) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 207) * The locking looks broadly like this:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 208) * mn_tree_invalidate_start(): mmu_interval_read_begin():
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 209) * spin_lock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 210) * seq = READ_ONCE(interval_sub->invalidate_seq);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 211) * seq == subs->invalidate_seq
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 212) * spin_unlock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 213) * spin_lock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 214) * seq = ++subscriptions->invalidate_seq
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 215) * spin_unlock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 216) * op->invalidate_range():
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 217) * user_lock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 218) * mmu_interval_set_seq()
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 219) * interval_sub->invalidate_seq = seq
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 220) * user_unlock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 221) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 222) * [Required: mmu_interval_read_retry() == true]
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 223) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 224) * mn_itree_inv_end():
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 225) * spin_lock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 226) * seq = ++subscriptions->invalidate_seq
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 227) * spin_unlock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 228) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 229) * user_lock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 230) * mmu_interval_read_retry():
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 231) * interval_sub->invalidate_seq != seq
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 232) * user_unlock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 233) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 234) * Barriers are not needed here as any races here are closed by an
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 235) * eventual mmu_interval_read_retry(), which provides a barrier via the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 236) * user_lock.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 237) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 238) spin_lock(&subscriptions->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 239) /* Pairs with the WRITE_ONCE in mmu_interval_set_seq() */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 240) seq = READ_ONCE(interval_sub->invalidate_seq);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 241) is_invalidating = seq == subscriptions->invalidate_seq;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 242) spin_unlock(&subscriptions->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 243)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 244) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 245) * interval_sub->invalidate_seq must always be set to an odd value via
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 246) * mmu_interval_set_seq() using the provided cur_seq from
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 247) * mn_itree_inv_start_range(). This ensures that if seq does wrap we
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 248) * will always clear the below sleep in some reasonable time as
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 249) * subscriptions->invalidate_seq is even in the idle state.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 250) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 251) lock_map_acquire(&__mmu_notifier_invalidate_range_start_map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 252) lock_map_release(&__mmu_notifier_invalidate_range_start_map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 253) if (is_invalidating)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 254) wait_event(subscriptions->wq,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 255) READ_ONCE(subscriptions->invalidate_seq) != seq);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 256)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 257) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 258) * Notice that mmu_interval_read_retry() can already be true at this
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 259) * point, avoiding loops here allows the caller to provide a global
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 260) * time bound.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 261) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 262)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 263) return seq;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 264) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 265) EXPORT_SYMBOL_GPL(mmu_interval_read_begin);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 266)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 267) static void mn_itree_release(struct mmu_notifier_subscriptions *subscriptions,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 268) struct mm_struct *mm)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 269) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 270) struct mmu_notifier_range range = {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 271) .flags = MMU_NOTIFIER_RANGE_BLOCKABLE,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 272) .event = MMU_NOTIFY_RELEASE,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 273) .mm = mm,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 274) .start = 0,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 275) .end = ULONG_MAX,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 276) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 277) struct mmu_interval_notifier *interval_sub;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 278) unsigned long cur_seq;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 279) bool ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 280)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 281) for (interval_sub =
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 282) mn_itree_inv_start_range(subscriptions, &range, &cur_seq);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 283) interval_sub;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 284) interval_sub = mn_itree_inv_next(interval_sub, &range)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 285) ret = interval_sub->ops->invalidate(interval_sub, &range,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 286) cur_seq);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 287) WARN_ON(!ret);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 288) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 289)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 290) mn_itree_inv_end(subscriptions);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 291) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 292)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 293) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 294) * This function can't run concurrently against mmu_notifier_register
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 295) * because mm->mm_users > 0 during mmu_notifier_register and exit_mmap
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 296) * runs with mm_users == 0. Other tasks may still invoke mmu notifiers
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 297) * in parallel despite there being no task using this mm any more,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 298) * through the vmas outside of the exit_mmap context, such as with
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 299) * vmtruncate. This serializes against mmu_notifier_unregister with
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 300) * the notifier_subscriptions->lock in addition to SRCU and it serializes
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 301) * against the other mmu notifiers with SRCU. struct mmu_notifier_subscriptions
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 302) * can't go away from under us as exit_mmap holds an mm_count pin
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 303) * itself.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 304) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 305) static void mn_hlist_release(struct mmu_notifier_subscriptions *subscriptions,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 306) struct mm_struct *mm)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 307) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 308) struct mmu_notifier *subscription;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 309) int id;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 310)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 311) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 312) * SRCU here will block mmu_notifier_unregister until
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 313) * ->release returns.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 314) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 315) id = srcu_read_lock(&srcu);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 316) hlist_for_each_entry_rcu(subscription, &subscriptions->list, hlist,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 317) srcu_read_lock_held(&srcu))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 318) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 319) * If ->release runs before mmu_notifier_unregister it must be
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 320) * handled, as it's the only way for the driver to flush all
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 321) * existing sptes and stop the driver from establishing any more
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 322) * sptes before all the pages in the mm are freed.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 323) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 324) if (subscription->ops->release)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 325) subscription->ops->release(subscription, mm);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 326)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 327) spin_lock(&subscriptions->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 328) while (unlikely(!hlist_empty(&subscriptions->list))) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 329) subscription = hlist_entry(subscriptions->list.first,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 330) struct mmu_notifier, hlist);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 331) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 332) * We arrived before mmu_notifier_unregister so
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 333) * mmu_notifier_unregister will do nothing other than to wait
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 334) * for ->release to finish and for mmu_notifier_unregister to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 335) * return.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 336) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 337) hlist_del_init_rcu(&subscription->hlist);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 338) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 339) spin_unlock(&subscriptions->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 340) srcu_read_unlock(&srcu, id);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 341)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 342) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 343) * synchronize_srcu here prevents mmu_notifier_release from returning to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 344) * exit_mmap (which would proceed with freeing all pages in the mm)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 345) * until the ->release method returns, if it was invoked by
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 346) * mmu_notifier_unregister.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 347) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 348) * The notifier_subscriptions can't go away from under us because
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 349) * one mm_count is held by exit_mmap.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 350) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 351) synchronize_srcu(&srcu);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 352) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 353)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 354) void __mmu_notifier_release(struct mm_struct *mm)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 355) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 356) struct mmu_notifier_subscriptions *subscriptions =
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 357) mm->notifier_subscriptions;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 358)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 359) if (subscriptions->has_itree)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 360) mn_itree_release(subscriptions, mm);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 361)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 362) if (!hlist_empty(&subscriptions->list))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 363) mn_hlist_release(subscriptions, mm);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 364) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 365)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 366) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 367) * If no young bitflag is supported by the hardware, ->clear_flush_young can
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 368) * unmap the address and return 1 or 0 depending if the mapping previously
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 369) * existed or not.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 370) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 371) int __mmu_notifier_clear_flush_young(struct mm_struct *mm,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 372) unsigned long start,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 373) unsigned long end)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 374) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 375) struct mmu_notifier *subscription;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 376) int young = 0, id;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 377)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 378) id = srcu_read_lock(&srcu);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 379) hlist_for_each_entry_rcu(subscription,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 380) &mm->notifier_subscriptions->list, hlist,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 381) srcu_read_lock_held(&srcu)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 382) if (subscription->ops->clear_flush_young)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 383) young |= subscription->ops->clear_flush_young(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 384) subscription, mm, start, end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 385) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 386) srcu_read_unlock(&srcu, id);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 387)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 388) return young;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 389) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 390)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 391) int __mmu_notifier_clear_young(struct mm_struct *mm,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 392) unsigned long start,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 393) unsigned long end)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 394) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 395) struct mmu_notifier *subscription;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 396) int young = 0, id;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 397)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 398) id = srcu_read_lock(&srcu);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 399) hlist_for_each_entry_rcu(subscription,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 400) &mm->notifier_subscriptions->list, hlist,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 401) srcu_read_lock_held(&srcu)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 402) if (subscription->ops->clear_young)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 403) young |= subscription->ops->clear_young(subscription,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 404) mm, start, end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 405) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 406) srcu_read_unlock(&srcu, id);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 407)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 408) return young;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 409) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 410)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 411) int __mmu_notifier_test_young(struct mm_struct *mm,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 412) unsigned long address)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 413) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 414) struct mmu_notifier *subscription;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 415) int young = 0, id;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 416)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 417) id = srcu_read_lock(&srcu);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 418) hlist_for_each_entry_rcu(subscription,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 419) &mm->notifier_subscriptions->list, hlist,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 420) srcu_read_lock_held(&srcu)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 421) if (subscription->ops->test_young) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 422) young = subscription->ops->test_young(subscription, mm,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 423) address);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 424) if (young)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 425) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 426) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 427) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 428) srcu_read_unlock(&srcu, id);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 429)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 430) return young;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 431) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 432)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 433) void __mmu_notifier_change_pte(struct mm_struct *mm, unsigned long address,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 434) pte_t pte)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 435) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 436) struct mmu_notifier *subscription;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 437) int id;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 438)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 439) id = srcu_read_lock(&srcu);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 440) hlist_for_each_entry_rcu(subscription,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 441) &mm->notifier_subscriptions->list, hlist,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 442) srcu_read_lock_held(&srcu)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 443) if (subscription->ops->change_pte)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 444) subscription->ops->change_pte(subscription, mm, address,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 445) pte);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 446) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 447) srcu_read_unlock(&srcu, id);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 448) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 449)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 450) static int mn_itree_invalidate(struct mmu_notifier_subscriptions *subscriptions,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 451) const struct mmu_notifier_range *range)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 452) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 453) struct mmu_interval_notifier *interval_sub;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 454) unsigned long cur_seq;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 455)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 456) for (interval_sub =
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 457) mn_itree_inv_start_range(subscriptions, range, &cur_seq);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 458) interval_sub;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 459) interval_sub = mn_itree_inv_next(interval_sub, range)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 460) bool ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 461)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 462) ret = interval_sub->ops->invalidate(interval_sub, range,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 463) cur_seq);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 464) if (!ret) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 465) if (WARN_ON(mmu_notifier_range_blockable(range)))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 466) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 467) goto out_would_block;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 468) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 469) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 470) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 471)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 472) out_would_block:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 473) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 474) * On -EAGAIN the non-blocking caller is not allowed to call
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 475) * invalidate_range_end()
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 476) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 477) mn_itree_inv_end(subscriptions);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 478) return -EAGAIN;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 479) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 480)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 481) static int mn_hlist_invalidate_range_start(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 482) struct mmu_notifier_subscriptions *subscriptions,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 483) struct mmu_notifier_range *range)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 484) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 485) struct mmu_notifier *subscription;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 486) int ret = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 487) int id;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 488)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 489) id = srcu_read_lock(&srcu);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 490) hlist_for_each_entry_rcu(subscription, &subscriptions->list, hlist,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 491) srcu_read_lock_held(&srcu)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 492) const struct mmu_notifier_ops *ops = subscription->ops;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 493)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 494) if (ops->invalidate_range_start) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 495) int _ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 496)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 497) if (!mmu_notifier_range_blockable(range))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 498) non_block_start();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 499) _ret = ops->invalidate_range_start(subscription, range);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 500) if (!mmu_notifier_range_blockable(range))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 501) non_block_end();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 502) if (_ret) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 503) pr_info("%pS callback failed with %d in %sblockable context.\n",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 504) ops->invalidate_range_start, _ret,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 505) !mmu_notifier_range_blockable(range) ?
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 506) "non-" :
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 507) "");
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 508) WARN_ON(mmu_notifier_range_blockable(range) ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 509) _ret != -EAGAIN);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 510) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 511) * We call all the notifiers on any EAGAIN,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 512) * there is no way for a notifier to know if
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 513) * its start method failed, thus a start that
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 514) * does EAGAIN can't also do end.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 515) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 516) WARN_ON(ops->invalidate_range_end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 517) ret = _ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 518) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 519) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 520) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 521)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 522) if (ret) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 523) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 524) * Must be non-blocking to get here. If there are multiple
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 525) * notifiers and one or more failed start, any that succeeded
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 526) * start are expecting their end to be called. Do so now.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 527) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 528) hlist_for_each_entry_rcu(subscription, &subscriptions->list,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 529) hlist, srcu_read_lock_held(&srcu)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 530) if (!subscription->ops->invalidate_range_end)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 531) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 532)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 533) subscription->ops->invalidate_range_end(subscription,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 534) range);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 535) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 536) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 537) srcu_read_unlock(&srcu, id);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 538)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 539) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 540) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 541)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 542) int __mmu_notifier_invalidate_range_start(struct mmu_notifier_range *range)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 543) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 544) struct mmu_notifier_subscriptions *subscriptions =
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 545) range->mm->notifier_subscriptions;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 546) int ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 547)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 548) if (subscriptions->has_itree) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 549) ret = mn_itree_invalidate(subscriptions, range);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 550) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 551) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 552) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 553) if (!hlist_empty(&subscriptions->list))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 554) return mn_hlist_invalidate_range_start(subscriptions, range);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 555) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 556) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 557)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 558) static void
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 559) mn_hlist_invalidate_end(struct mmu_notifier_subscriptions *subscriptions,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 560) struct mmu_notifier_range *range, bool only_end)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 561) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 562) struct mmu_notifier *subscription;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 563) int id;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 564)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 565) id = srcu_read_lock(&srcu);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 566) hlist_for_each_entry_rcu(subscription, &subscriptions->list, hlist,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 567) srcu_read_lock_held(&srcu)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 568) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 569) * Call invalidate_range here too to avoid the need for the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 570) * subsystem of having to register an invalidate_range_end
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 571) * call-back when there is invalidate_range already. Usually a
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 572) * subsystem registers either invalidate_range_start()/end() or
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 573) * invalidate_range(), so this will be no additional overhead
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 574) * (besides the pointer check).
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 575) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 576) * We skip call to invalidate_range() if we know it is safe ie
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 577) * call site use mmu_notifier_invalidate_range_only_end() which
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 578) * is safe to do when we know that a call to invalidate_range()
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 579) * already happen under page table lock.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 580) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 581) if (!only_end && subscription->ops->invalidate_range)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 582) subscription->ops->invalidate_range(subscription,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 583) range->mm,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 584) range->start,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 585) range->end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 586) if (subscription->ops->invalidate_range_end) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 587) if (!mmu_notifier_range_blockable(range))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 588) non_block_start();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 589) subscription->ops->invalidate_range_end(subscription,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 590) range);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 591) if (!mmu_notifier_range_blockable(range))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 592) non_block_end();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 593) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 594) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 595) srcu_read_unlock(&srcu, id);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 596) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 597)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 598) void __mmu_notifier_invalidate_range_end(struct mmu_notifier_range *range,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 599) bool only_end)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 600) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 601) struct mmu_notifier_subscriptions *subscriptions =
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 602) range->mm->notifier_subscriptions;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 603)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 604) lock_map_acquire(&__mmu_notifier_invalidate_range_start_map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 605) if (subscriptions->has_itree)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 606) mn_itree_inv_end(subscriptions);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 607)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 608) if (!hlist_empty(&subscriptions->list))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 609) mn_hlist_invalidate_end(subscriptions, range, only_end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 610) lock_map_release(&__mmu_notifier_invalidate_range_start_map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 611) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 612)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 613) void __mmu_notifier_invalidate_range(struct mm_struct *mm,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 614) unsigned long start, unsigned long end)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 615) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 616) struct mmu_notifier *subscription;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 617) int id;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 618)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 619) id = srcu_read_lock(&srcu);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 620) hlist_for_each_entry_rcu(subscription,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 621) &mm->notifier_subscriptions->list, hlist,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 622) srcu_read_lock_held(&srcu)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 623) if (subscription->ops->invalidate_range)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 624) subscription->ops->invalidate_range(subscription, mm,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 625) start, end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 626) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 627) srcu_read_unlock(&srcu, id);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 628) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 629)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 630) #ifdef CONFIG_SPECULATIVE_PAGE_FAULT
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 631)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 632) static inline void mmu_notifier_write_lock(struct mm_struct *mm)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 633) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 634) percpu_down_write(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 635) &mm->notifier_subscriptions->hdr.mmu_notifier_lock->rw_sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 636) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 637)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 638) static inline void mmu_notifier_write_unlock(struct mm_struct *mm)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 639) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 640) percpu_up_write(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 641) &mm->notifier_subscriptions->hdr.mmu_notifier_lock->rw_sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 642) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 643)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 644) #else /* CONFIG_SPECULATIVE_PAGE_FAULT */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 645)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 646) static inline void mmu_notifier_write_lock(struct mm_struct *mm) {}
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 647) static inline void mmu_notifier_write_unlock(struct mm_struct *mm) {}
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 648)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 649) #endif /* CONFIG_SPECULATIVE_PAGE_FAULT */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 650)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 651) static void init_subscriptions(struct mmu_notifier_subscriptions *subscriptions)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 652) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 653) INIT_HLIST_HEAD(&subscriptions->list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 654) spin_lock_init(&subscriptions->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 655) subscriptions->invalidate_seq = 2;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 656) subscriptions->itree = RB_ROOT_CACHED;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 657) init_waitqueue_head(&subscriptions->wq);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 658) INIT_HLIST_HEAD(&subscriptions->deferred_list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 659) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 660)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 661) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 662) * Same as mmu_notifier_register but here the caller must hold the mmap_lock in
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 663) * write mode. A NULL mn signals the notifier is being registered for itree
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 664) * mode.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 665) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 666) int __mmu_notifier_register(struct mmu_notifier *subscription,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 667) struct mm_struct *mm)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 668) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 669) struct mmu_notifier_subscriptions *subscriptions = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 670) int ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 671)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 672) mmap_assert_write_locked(mm);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 673) BUG_ON(atomic_read(&mm->mm_users) <= 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 674)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 675) if (IS_ENABLED(CONFIG_LOCKDEP)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 676) fs_reclaim_acquire(GFP_KERNEL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 677) lock_map_acquire(&__mmu_notifier_invalidate_range_start_map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 678) lock_map_release(&__mmu_notifier_invalidate_range_start_map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 679) fs_reclaim_release(GFP_KERNEL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 680) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 681)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 682) if (!mm->notifier_subscriptions) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 683) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 684) * kmalloc cannot be called under mm_take_all_locks(), but we
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 685) * know that mm->notifier_subscriptions can't change while we
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 686) * hold the write side of the mmap_lock.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 687) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 688) subscriptions = kzalloc(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 689) sizeof(struct mmu_notifier_subscriptions), GFP_KERNEL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 690) if (!subscriptions)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 691) return -ENOMEM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 692)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 693) init_subscriptions(subscriptions);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 694) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 695)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 696) mmu_notifier_write_lock(mm);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 697)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 698) ret = mm_take_all_locks(mm);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 699) if (unlikely(ret)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 700) mmu_notifier_write_unlock(mm);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 701) goto out_clean;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 702) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 703)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 704) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 705) * Serialize the update against mmu_notifier_unregister. A
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 706) * side note: mmu_notifier_release can't run concurrently with
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 707) * us because we hold the mm_users pin (either implicitly as
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 708) * current->mm or explicitly with get_task_mm() or similar).
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 709) * We can't race against any other mmu notifier method either
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 710) * thanks to mm_take_all_locks().
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 711) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 712) * release semantics on the initialization of the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 713) * mmu_notifier_subscriptions's contents are provided for unlocked
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 714) * readers. acquire can only be used while holding the mmgrab or
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 715) * mmget, and is safe because once created the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 716) * mmu_notifier_subscriptions is not freed until the mm is destroyed.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 717) * As above, users holding the mmap_lock or one of the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 718) * mm_take_all_locks() do not need to use acquire semantics.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 719) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 720) if (subscriptions)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 721) smp_store_release(&mm->notifier_subscriptions, subscriptions);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 722) mm->notifier_subscriptions->hdr.valid = true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 723)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 724) if (subscription) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 725) /* Pairs with the mmdrop in mmu_notifier_unregister_* */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 726) mmgrab(mm);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 727) subscription->mm = mm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 728) subscription->users = 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 729)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 730) spin_lock(&mm->notifier_subscriptions->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 731) hlist_add_head_rcu(&subscription->hlist,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 732) &mm->notifier_subscriptions->list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 733) spin_unlock(&mm->notifier_subscriptions->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 734) } else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 735) mm->notifier_subscriptions->has_itree = true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 736)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 737) mm_drop_all_locks(mm);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 738) mmu_notifier_write_unlock(mm);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 739) BUG_ON(atomic_read(&mm->mm_users) <= 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 740) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 741)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 742) out_clean:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 743) kfree(subscriptions);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 744) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 745) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 746) EXPORT_SYMBOL_GPL(__mmu_notifier_register);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 747)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 748) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 749) * mmu_notifier_register - Register a notifier on a mm
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 750) * @subscription: The notifier to attach
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 751) * @mm: The mm to attach the notifier to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 752) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 753) * Must not hold mmap_lock nor any other VM related lock when calling
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 754) * this registration function. Must also ensure mm_users can't go down
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 755) * to zero while this runs to avoid races with mmu_notifier_release,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 756) * so mm has to be current->mm or the mm should be pinned safely such
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 757) * as with get_task_mm(). If the mm is not current->mm, the mm_users
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 758) * pin should be released by calling mmput after mmu_notifier_register
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 759) * returns.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 760) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 761) * mmu_notifier_unregister() or mmu_notifier_put() must be always called to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 762) * unregister the notifier.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 763) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 764) * While the caller has a mmu_notifier get the subscription->mm pointer will remain
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 765) * valid, and can be converted to an active mm pointer via mmget_not_zero().
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 766) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 767) int mmu_notifier_register(struct mmu_notifier *subscription,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 768) struct mm_struct *mm)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 769) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 770) int ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 771)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 772) mmap_write_lock(mm);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 773) ret = __mmu_notifier_register(subscription, mm);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 774) mmap_write_unlock(mm);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 775) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 776) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 777) EXPORT_SYMBOL_GPL(mmu_notifier_register);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 778)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 779) static struct mmu_notifier *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 780) find_get_mmu_notifier(struct mm_struct *mm, const struct mmu_notifier_ops *ops)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 781) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 782) struct mmu_notifier *subscription;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 783)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 784) spin_lock(&mm->notifier_subscriptions->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 785) hlist_for_each_entry_rcu(subscription,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 786) &mm->notifier_subscriptions->list, hlist,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 787) lockdep_is_held(&mm->notifier_subscriptions->lock)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 788) if (subscription->ops != ops)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 789) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 790)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 791) if (likely(subscription->users != UINT_MAX))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 792) subscription->users++;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 793) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 794) subscription = ERR_PTR(-EOVERFLOW);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 795) spin_unlock(&mm->notifier_subscriptions->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 796) return subscription;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 797) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 798) spin_unlock(&mm->notifier_subscriptions->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 799) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 800) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 801)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 802) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 803) * mmu_notifier_get_locked - Return the single struct mmu_notifier for
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 804) * the mm & ops
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 805) * @ops: The operations struct being subscribe with
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 806) * @mm : The mm to attach notifiers too
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 807) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 808) * This function either allocates a new mmu_notifier via
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 809) * ops->alloc_notifier(), or returns an already existing notifier on the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 810) * list. The value of the ops pointer is used to determine when two notifiers
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 811) * are the same.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 812) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 813) * Each call to mmu_notifier_get() must be paired with a call to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 814) * mmu_notifier_put(). The caller must hold the write side of mm->mmap_lock.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 815) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 816) * While the caller has a mmu_notifier get the mm pointer will remain valid,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 817) * and can be converted to an active mm pointer via mmget_not_zero().
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 818) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 819) struct mmu_notifier *mmu_notifier_get_locked(const struct mmu_notifier_ops *ops,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 820) struct mm_struct *mm)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 821) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 822) struct mmu_notifier *subscription;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 823) int ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 824)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 825) mmap_assert_write_locked(mm);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 826)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 827) if (mm->notifier_subscriptions) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 828) subscription = find_get_mmu_notifier(mm, ops);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 829) if (subscription)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 830) return subscription;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 831) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 832)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 833) subscription = ops->alloc_notifier(mm);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 834) if (IS_ERR(subscription))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 835) return subscription;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 836) subscription->ops = ops;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 837) ret = __mmu_notifier_register(subscription, mm);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 838) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 839) goto out_free;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 840) return subscription;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 841) out_free:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 842) subscription->ops->free_notifier(subscription);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 843) return ERR_PTR(ret);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 844) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 845) EXPORT_SYMBOL_GPL(mmu_notifier_get_locked);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 846)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 847) /* this is called after the last mmu_notifier_unregister() returned */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 848) void __mmu_notifier_subscriptions_destroy(struct mm_struct *mm)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 849) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 850) BUG_ON(!hlist_empty(&mm->notifier_subscriptions->list));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 851) kfree(mm->notifier_subscriptions);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 852) mm->notifier_subscriptions = LIST_POISON1; /* debug */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 853) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 854)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 855) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 856) * This releases the mm_count pin automatically and frees the mm
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 857) * structure if it was the last user of it. It serializes against
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 858) * running mmu notifiers with SRCU and against mmu_notifier_unregister
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 859) * with the unregister lock + SRCU. All sptes must be dropped before
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 860) * calling mmu_notifier_unregister. ->release or any other notifier
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 861) * method may be invoked concurrently with mmu_notifier_unregister,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 862) * and only after mmu_notifier_unregister returned we're guaranteed
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 863) * that ->release or any other method can't run anymore.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 864) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 865) void mmu_notifier_unregister(struct mmu_notifier *subscription,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 866) struct mm_struct *mm)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 867) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 868) BUG_ON(atomic_read(&mm->mm_count) <= 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 869)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 870) if (!hlist_unhashed(&subscription->hlist)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 871) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 872) * SRCU here will force exit_mmap to wait for ->release to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 873) * finish before freeing the pages.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 874) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 875) int id;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 876)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 877) id = srcu_read_lock(&srcu);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 878) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 879) * exit_mmap will block in mmu_notifier_release to guarantee
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 880) * that ->release is called before freeing the pages.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 881) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 882) if (subscription->ops->release)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 883) subscription->ops->release(subscription, mm);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 884) srcu_read_unlock(&srcu, id);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 885)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 886) spin_lock(&mm->notifier_subscriptions->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 887) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 888) * Can not use list_del_rcu() since __mmu_notifier_release
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 889) * can delete it before we hold the lock.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 890) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 891) hlist_del_init_rcu(&subscription->hlist);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 892) spin_unlock(&mm->notifier_subscriptions->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 893) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 894)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 895) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 896) * Wait for any running method to finish, of course including
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 897) * ->release if it was run by mmu_notifier_release instead of us.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 898) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 899) synchronize_srcu(&srcu);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 900)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 901) BUG_ON(atomic_read(&mm->mm_count) <= 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 902)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 903) mmdrop(mm);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 904) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 905) EXPORT_SYMBOL_GPL(mmu_notifier_unregister);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 906)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 907) static void mmu_notifier_free_rcu(struct rcu_head *rcu)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 908) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 909) struct mmu_notifier *subscription =
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 910) container_of(rcu, struct mmu_notifier, rcu);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 911) struct mm_struct *mm = subscription->mm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 912)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 913) subscription->ops->free_notifier(subscription);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 914) /* Pairs with the get in __mmu_notifier_register() */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 915) mmdrop(mm);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 916) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 917)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 918) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 919) * mmu_notifier_put - Release the reference on the notifier
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 920) * @subscription: The notifier to act on
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 921) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 922) * This function must be paired with each mmu_notifier_get(), it releases the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 923) * reference obtained by the get. If this is the last reference then process
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 924) * to free the notifier will be run asynchronously.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 925) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 926) * Unlike mmu_notifier_unregister() the get/put flow only calls ops->release
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 927) * when the mm_struct is destroyed. Instead free_notifier is always called to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 928) * release any resources held by the user.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 929) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 930) * As ops->release is not guaranteed to be called, the user must ensure that
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 931) * all sptes are dropped, and no new sptes can be established before
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 932) * mmu_notifier_put() is called.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 933) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 934) * This function can be called from the ops->release callback, however the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 935) * caller must still ensure it is called pairwise with mmu_notifier_get().
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 936) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 937) * Modules calling this function must call mmu_notifier_synchronize() in
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 938) * their __exit functions to ensure the async work is completed.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 939) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 940) void mmu_notifier_put(struct mmu_notifier *subscription)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 941) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 942) struct mm_struct *mm = subscription->mm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 943)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 944) spin_lock(&mm->notifier_subscriptions->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 945) if (WARN_ON(!subscription->users) || --subscription->users)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 946) goto out_unlock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 947) hlist_del_init_rcu(&subscription->hlist);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 948) spin_unlock(&mm->notifier_subscriptions->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 949)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 950) call_srcu(&srcu, &subscription->rcu, mmu_notifier_free_rcu);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 951) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 952)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 953) out_unlock:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 954) spin_unlock(&mm->notifier_subscriptions->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 955) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 956) EXPORT_SYMBOL_GPL(mmu_notifier_put);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 957)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 958) static int __mmu_interval_notifier_insert(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 959) struct mmu_interval_notifier *interval_sub, struct mm_struct *mm,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 960) struct mmu_notifier_subscriptions *subscriptions, unsigned long start,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 961) unsigned long length, const struct mmu_interval_notifier_ops *ops)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 962) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 963) interval_sub->mm = mm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 964) interval_sub->ops = ops;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 965) RB_CLEAR_NODE(&interval_sub->interval_tree.rb);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 966) interval_sub->interval_tree.start = start;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 967) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 968) * Note that the representation of the intervals in the interval tree
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 969) * considers the ending point as contained in the interval.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 970) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 971) if (length == 0 ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 972) check_add_overflow(start, length - 1,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 973) &interval_sub->interval_tree.last))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 974) return -EOVERFLOW;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 975)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 976) /* Must call with a mmget() held */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 977) if (WARN_ON(atomic_read(&mm->mm_users) <= 0))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 978) return -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 979)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 980) /* pairs with mmdrop in mmu_interval_notifier_remove() */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 981) mmgrab(mm);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 982)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 983) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 984) * If some invalidate_range_start/end region is going on in parallel
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 985) * we don't know what VA ranges are affected, so we must assume this
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 986) * new range is included.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 987) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 988) * If the itree is invalidating then we are not allowed to change
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 989) * it. Retrying until invalidation is done is tricky due to the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 990) * possibility for live lock, instead defer the add to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 991) * mn_itree_inv_end() so this algorithm is deterministic.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 992) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 993) * In all cases the value for the interval_sub->invalidate_seq should be
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 994) * odd, see mmu_interval_read_begin()
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 995) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 996) spin_lock(&subscriptions->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 997) if (subscriptions->active_invalidate_ranges) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 998) if (mn_itree_is_invalidating(subscriptions))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 999) hlist_add_head(&interval_sub->deferred_item,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1000) &subscriptions->deferred_list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1001) else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1002) subscriptions->invalidate_seq |= 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1003) interval_tree_insert(&interval_sub->interval_tree,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1004) &subscriptions->itree);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1005) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1006) interval_sub->invalidate_seq = subscriptions->invalidate_seq;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1007) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1008) WARN_ON(mn_itree_is_invalidating(subscriptions));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1009) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1010) * The starting seq for a subscription not under invalidation
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1011) * should be odd, not equal to the current invalidate_seq and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1012) * invalidate_seq should not 'wrap' to the new seq any time
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1013) * soon.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1014) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1015) interval_sub->invalidate_seq =
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1016) subscriptions->invalidate_seq - 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1017) interval_tree_insert(&interval_sub->interval_tree,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1018) &subscriptions->itree);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1019) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1020) spin_unlock(&subscriptions->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1021) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1022) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1023)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1024) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1025) * mmu_interval_notifier_insert - Insert an interval notifier
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1026) * @interval_sub: Interval subscription to register
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1027) * @start: Starting virtual address to monitor
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1028) * @length: Length of the range to monitor
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1029) * @mm: mm_struct to attach to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1030) * @ops: Interval notifier operations to be called on matching events
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1031) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1032) * This function subscribes the interval notifier for notifications from the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1033) * mm. Upon return the ops related to mmu_interval_notifier will be called
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1034) * whenever an event that intersects with the given range occurs.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1035) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1036) * Upon return the range_notifier may not be present in the interval tree yet.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1037) * The caller must use the normal interval notifier read flow via
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1038) * mmu_interval_read_begin() to establish SPTEs for this range.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1039) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1040) int mmu_interval_notifier_insert(struct mmu_interval_notifier *interval_sub,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1041) struct mm_struct *mm, unsigned long start,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1042) unsigned long length,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1043) const struct mmu_interval_notifier_ops *ops)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1044) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1045) struct mmu_notifier_subscriptions *subscriptions;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1046) int ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1047)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1048) might_lock(&mm->mmap_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1049)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1050) subscriptions = smp_load_acquire(&mm->notifier_subscriptions);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1051) if (!subscriptions || !subscriptions->has_itree) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1052) ret = mmu_notifier_register(NULL, mm);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1053) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1054) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1055) subscriptions = mm->notifier_subscriptions;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1056) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1057) return __mmu_interval_notifier_insert(interval_sub, mm, subscriptions,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1058) start, length, ops);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1059) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1060) EXPORT_SYMBOL_GPL(mmu_interval_notifier_insert);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1061)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1062) int mmu_interval_notifier_insert_locked(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1063) struct mmu_interval_notifier *interval_sub, struct mm_struct *mm,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1064) unsigned long start, unsigned long length,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1065) const struct mmu_interval_notifier_ops *ops)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1066) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1067) struct mmu_notifier_subscriptions *subscriptions =
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1068) mm->notifier_subscriptions;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1069) int ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1070)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1071) mmap_assert_write_locked(mm);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1072)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1073) if (!subscriptions || !subscriptions->has_itree) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1074) ret = __mmu_notifier_register(NULL, mm);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1075) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1076) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1077) subscriptions = mm->notifier_subscriptions;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1078) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1079) return __mmu_interval_notifier_insert(interval_sub, mm, subscriptions,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1080) start, length, ops);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1081) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1082) EXPORT_SYMBOL_GPL(mmu_interval_notifier_insert_locked);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1083)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1084) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1085) * mmu_interval_notifier_remove - Remove a interval notifier
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1086) * @interval_sub: Interval subscription to unregister
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1087) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1088) * This function must be paired with mmu_interval_notifier_insert(). It cannot
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1089) * be called from any ops callback.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1090) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1091) * Once this returns ops callbacks are no longer running on other CPUs and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1092) * will not be called in future.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1093) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1094) void mmu_interval_notifier_remove(struct mmu_interval_notifier *interval_sub)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1095) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1096) struct mm_struct *mm = interval_sub->mm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1097) struct mmu_notifier_subscriptions *subscriptions =
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1098) mm->notifier_subscriptions;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1099) unsigned long seq = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1100)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1101) might_sleep();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1102)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1103) spin_lock(&subscriptions->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1104) if (mn_itree_is_invalidating(subscriptions)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1105) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1106) * remove is being called after insert put this on the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1107) * deferred list, but before the deferred list was processed.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1108) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1109) if (RB_EMPTY_NODE(&interval_sub->interval_tree.rb)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1110) hlist_del(&interval_sub->deferred_item);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1111) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1112) hlist_add_head(&interval_sub->deferred_item,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1113) &subscriptions->deferred_list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1114) seq = subscriptions->invalidate_seq;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1115) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1116) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1117) WARN_ON(RB_EMPTY_NODE(&interval_sub->interval_tree.rb));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1118) interval_tree_remove(&interval_sub->interval_tree,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1119) &subscriptions->itree);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1120) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1121) spin_unlock(&subscriptions->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1122)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1123) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1124) * The possible sleep on progress in the invalidation requires the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1125) * caller not hold any locks held by invalidation callbacks.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1126) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1127) lock_map_acquire(&__mmu_notifier_invalidate_range_start_map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1128) lock_map_release(&__mmu_notifier_invalidate_range_start_map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1129) if (seq)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1130) wait_event(subscriptions->wq,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1131) READ_ONCE(subscriptions->invalidate_seq) != seq);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1132)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1133) /* pairs with mmgrab in mmu_interval_notifier_insert() */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1134) mmdrop(mm);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1135) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1136) EXPORT_SYMBOL_GPL(mmu_interval_notifier_remove);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1137)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1138) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1139) * mmu_notifier_synchronize - Ensure all mmu_notifiers are freed
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1140) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1141) * This function ensures that all outstanding async SRU work from
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1142) * mmu_notifier_put() is completed. After it returns any mmu_notifier_ops
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1143) * associated with an unused mmu_notifier will no longer be called.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1144) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1145) * Before using the caller must ensure that all of its mmu_notifiers have been
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1146) * fully released via mmu_notifier_put().
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1147) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1148) * Modules using the mmu_notifier_put() API should call this in their __exit
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1149) * function to avoid module unloading races.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1150) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1151) void mmu_notifier_synchronize(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1152) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1153) synchronize_srcu(&srcu);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1154) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1155) EXPORT_SYMBOL_GPL(mmu_notifier_synchronize);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1156)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1157) bool
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1158) mmu_notifier_range_update_to_read_only(const struct mmu_notifier_range *range)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1159) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1160) if (!range->vma || range->event != MMU_NOTIFY_PROTECTION_VMA)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1161) return false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1162) /* Return true if the vma still have the read flag set. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1163) return range->vma->vm_flags & VM_READ;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1164) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1165) EXPORT_SYMBOL_GPL(mmu_notifier_range_update_to_read_only);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1166)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1167) #ifdef CONFIG_SPECULATIVE_PAGE_FAULT
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1168)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1169) bool mmu_notifier_subscriptions_init(struct mm_struct *mm)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1170) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1171) struct mmu_notifier_subscriptions *subscriptions;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1172) struct percpu_rw_semaphore_atomic *sem;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1173)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1174) subscriptions = kzalloc(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1175) sizeof(struct mmu_notifier_subscriptions), GFP_KERNEL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1176) if (!subscriptions)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1177) return false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1178)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1179) sem = kzalloc(sizeof(struct percpu_rw_semaphore_atomic), GFP_KERNEL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1180) if (!sem) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1181) kfree(subscriptions);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1182) return false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1183) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1184) percpu_init_rwsem(&sem->rw_sem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1185)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1186) init_subscriptions(subscriptions);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1187) subscriptions->has_itree = true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1188) subscriptions->hdr.valid = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1189) subscriptions->hdr.mmu_notifier_lock = sem;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1190) mm->notifier_subscriptions = subscriptions;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1191)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1192) return true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1193) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1194)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1195) void mmu_notifier_subscriptions_destroy(struct mm_struct *mm)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1196) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1197) percpu_rwsem_async_destroy(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1198) mm->notifier_subscriptions->hdr.mmu_notifier_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1199) kfree(mm->notifier_subscriptions);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1200) mm->notifier_subscriptions = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1201) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1202)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1203) #endif /* CONFIG_SPECULATIVE_PAGE_FAULT */
Orange Pi5 kernel
Deprecated Linux kernel 5.10.110 for OrangePi 5/5B/5+ boards
3 Commits
0 Branches
0 Tags